Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling

Page created by Max Cohen
 
CONTINUE READING
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489
                                    https://doi.org/10.1038/s41419-021-03788-4                                                                                                             Cell Death & Disease

                                     ARTICLE                                                                                                                                                    Open Access

                                    Monocytic MDSC mobilization promotes tumor
                                    recurrence after liver transplantation via CXCL10/
                                    TLR4/MMP14 signaling
                                    Hui Liu1, Chang Chun Ling1,2, Wai Ho Oscar Yeung1, Li Pang1, Jiang Liu1, Jie Zhou1, Wei Yi Zhang1, Xiao Bing Liu1,
                                    Tak Pan Kevin Ng 1, Xin Xiang Yang1, Chung Mau Lo1 and Kwan Man1

                                      Abstract
                                      Tumor recurrence is the major obstacle for pushing the envelope of liver transplantation for hepatocellular carcinoma
                                      (HCC) patients. The inflammatory cascades activated by acute liver graft injury promote tumor recurrence. We aimed to
                                      explore the role and mechanism of myeloid-derived suppressor cell (MDSC) mobilization induced by liver graft injury on
                                      tumor recurrence. By analyzing 331 HCC patients who received liver transplantation, the patients with graft weight ratio
                                      (GWR, the weight of liver graft divided by the estimated standard liver weight of recipient)
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489                                                                  Page 2 of 14

infiltrated into the liver in hepatic diseases8,9. In the        analysis. Among them, 50 patients experienced HCC
development of hepatic ischemia/reperfusion injury, it has      recurrence (recurrence) while the other 281 patients had
been identified that CXCL10 is required for the induction        no HCC recurrence (non-recurrence). The graft weight
of pro-inflammatory responses10. Oncologically, CXCL10,          ratio (%) (GWR) has been defined as the weight of liver
with its receptor C-X-C motif chemokine receptor 3              graft divided by the estimated standard liver weight of the
(CXCR3), facilitates cancer cell proliferation, metastasis,     recipient25. According to the GWR, the patients were
and invasion11–13. CXCL10 also contributes to inflam-            assigned as large liver graft group (GWR ≥60%, n = 149)
matory responses or hepatocellular apoptosis through            and small-for-size liver graft group (GWR < 60%, n =
toll-like receptor 4 (TLR4), but not CXCR314,15. The            180). The liver graft biopsies and blood were collected at
intestinal venous congestion caused by liver graft injury       2 h and 7 days after portal vein reperfusion, respectively.
facilitated cancer recurrence by TLR4 increase in rodent        All clinical record access and tissue sample collection
models16. However, few studies reported the mechanism           were approved with signed consent forms by each donor
of CXCL10/TLR4 signaling regulating immune cells for            and recipient. The procedures followed in this study was
promoting liver cancer recurrence.                              in accordance with the ethical standards of the Institu-
   A key cell population defined as myeloid-derived sup-         tional Review Board (IRB) of The University of Hong
pressor cells (MDSCs) contributes to the immunosup-             Kong and with the Helsinki declaration of 1975, as revised
pression in tumors, which is critical in tumor evasion of       in 1983. No donor organs were obtained from executed
the immune system17. MDSC can disrupt immune sur-               prisoners or other institutionalized persons.
veillance by suppressing effector T cells18, blocking nature
killer cell cytotoxicity19, expanding regulatory T cells20,     Animal models
and skewing macrophages into immunosuppressive M2                  Male Sprague Dawley (SD), Buffalo rats, and C57 BL/6
phenotype21. The liver is a favorable site for MDSC             mice (6–8 weeks) were purchased from the Laboratory
expansion17. Recent studies indicated that the hypoxia          Animal Unit, The University of Hong Kong. The
environment recruited MDSCs to promote HCC22,23.                CXCL10, CXCR3, and TLR4 knockout mice (6–8 weeks)
Chemokines modulate the hepatic microenvironment and            were described in previous papers with the controls
regulate critical aspects in the pathogenesis of liver can-     cohoused26,27. All animals were housed in a standard
cer24. However, the relationship of chemokines, especially      animal facility at 22 ± 2 °C under controlled 12-h light/
CXCL10 signals, with MDSC mobilization in HCC                   dark cycles. Both rats and mice had free access to regular
recurrence after transplantation has never been explored.       chow (5053-PicoLab®Rodent Diet 20, Lab Diet, MO,
   In this study, we aimed to explore the role and              USA) and autoclaved water. According to our preliminary
mechanism of MDSC mobilization by CXCL10 signaling              experiment, five animals were chosen in each group (N =
at early graft injury, which might lead to tumor recurrence     5). All the animals were randomly grouping with no
post-liver transplantation. We demonstrated the associa-        blinding. The animals with poor physical conditions were
tion among acute-phase MDSC recruitment, CXCL10/                excluded. All animals received humane care according to
TLR4 overexpression, and tumor recurrence after small-          the criteria outlined in Guide for the Care and Use of
for-size liver graft transplantation in both clinical and rat   Laboratory Animals (National Institutes Health publica-
transplant studies. Furthermore, we provided new insights       tion 86–23, 1985 revision). The experimental protocols
that CXCL10 through TLR4, instead of CXCR3, mobi-               were performed under the guidelines approved by the
lized monocytic MDSCs, but not granulocytic MDSCs, to           Committee on the Use of Live Animals in Teaching and
the liver for facilitating cancer recurrence using              Research, The University of Hong Kong
CXCL10−/−, CXCR3−/− and TLR4−/− mice. Higher TLR4                  (1) Rat orthotopic liver transplantation model
expression on monocytic MDSCs might explain why the                The orthotopic liver transplantation model was estab-
monocytic MDSCs were mobilized. The direct role of              lished using whole graft and small-for-size graft (ratio of
CXCL10 on MDSC mobilization through TLR4 was ver-               graft weight to recipient liver weight was about 50%) in
ified through in vitro studies. We also discovered a             SD rats. Blood and liver tissues were collected on Day 1, 3,
motility-related molecule, matrix metallopeptidase 14           5 after transplantation to investigate MDSC mobilization
(MMP14), bridged CXCL10/TLR4 signaling and MDSC                 and CXCL10/TLR4 levels in early-stage liver graft injury.
mobilization.                                                   The management of rat orthotopic liver transplantation
                                                                was implemented according to the surgical protocol
Materials and methods                                           described previously28,29. To analyze the MDSC dis-
Clinical cohort and biopsies                                    tribution, CXCL10/TLR4 expressions, and intragraft
 From 1995 to 2016, 331 patients with HCC who                   angiogenesis in recurrent tumors, a rat orthotopic liver
underwent liver transplantation in Queen Mary Hospital,         transplantation with tumor recurrence model in Buffalo
The University of Hong Kong, were included in our               rats were used as described in our previous papers30,31.

Official journal of the Cell Death Differentiation Association
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489                                                                   Page 3 of 14

The tumor and non-tumor tissues of the liver graft were         and Graphpad Prism 5.0 (GraphPad Software Inc,
collected at day 14 post-transplantation.                       CA, USA).
  (2) Mouse hepatic ischemia/reperfusion plus major
hepatectomy (IRH) model                                         Results
  To mimic liver transplantation using a small-for-size         Tumor recurrent rate was higher in HCC patients after
graft, the wild type (C57 BL/6), CXCL10−/−, CXCR3−/−,           small-for-size liver graft transplantation
and TLR4−/− mice were subjected to partial ischemia/              Three hundred and thirty-one HCC patients who
reperfusion injury plus major hepatectomy26. The right          received liver transplantation were recruited in this study.
branch of the portal vein and hepatic arterial was clamped      HCC recurrence occurred in 50 patients post-liver
for 45 min, meanwhile, the left and caudate lobes were          transplantation. According to the analysis of clinical
removed before reperfusion. Blood and liver tissues were        parameters, the recurrence was more frequent in the
collected on Day 1, 3, 5 after reperfusion.                     patients with positive HBsAg (p = 0.01), higher serum
  The mouse liver tumor cells (Hepa1-6, ATCC, 2 × 106/          alpha-fetoprotein (AFP) (p = 0.000), more number of
100 μl per mouse) were injected into the portal vein            tumors (p = 0.002), the larger size of tumor largest size
immediately after hepatic IRH to establish mouse liver          (p = 0.003), macrovascular invasion (p = 0.000), advanced
IRH with tumor recurrence model. To investigate the role        TNM stage (p = 0.000), beyond Milan criteria (p = 0.000)
of TLR4 on MDSC mobilization in liver tumor recur-              and UCSF criteria (p = 0.000), vascular permeation (p =
rence, CLI095 (TLR4 inhibitor, 3 mg/kg, Invivogen, CA,          0.000) and higher differentiation of tumor (p = 0.025).
USA) was injected intraperitoneally 2 h before the              Importantly, tumor recurrent rate was significantly higher
operation and every two days after the reperfusion. About       in the HCC patients with small-for-size graft (GWR <
21 days after the surgical process, blood and liver tissues     60%) post-liver transplantation (p = 0.018) (Table 1).
were collected for analysis.                                    Consistent with the results, the survival analysis demon-
                                                                strated that the patients with GWR < 60% had poor
Primary MDSC isolation by magnetic bead cell sorting            disease-free survival compared to patients with GWR ≥
from mouse spleen and bone marrow                               60%, while no significant difference in overall survival
  Bone marrow cells were obtained from femurs and tibias        after liver transplantation (Fig. 1A).
of mice. The splenic cells and bone marrow cells were
flushed out using phosphate-buffered saline by syringe.          MDSCs were increased in HCC patients with small-for-size
The cell suspensions were filtered through 70 μm cell            graft or tumor recurrence post-liver transplantation
strainers and treated with ACK lysing buffer (Chem Cruz,          More circulatory (Day 7, p = 0.0384) and intragraft (2 h
Santa Cruz Biotechnology, TX, USA). After obtaining the         post perfusion, p = 0.0076) MDSCs were found in the
single-cell suspension, Gr1+CD11b+ cells (MDSCs) were           patients with small-for-size graft compared to large graft
isolated using a mouse MDSC isolation kit (Stemcell             post-liver transplantation (Fig. 1B). Furthermore, the
Technologies, BC, Canada), according to the manu-               patients with HCC recurrence had a higher population of
facturer’s instruction. The MDSC purity was evaluated by        the circulatory (Day 7, p = 0.0062) and intragraft (2 h post
Gr1+CD11b+ (Stemcell Technologies; BD Pharmingen,               perfusion, p = 0.0451) MDSCs than non-recurrent
CA, USA) population >90% via flow cytometry.                     patients after transplantation (Fig. 1C). The patients
                                                                with high MDSCs had poor disease-free survival (Fig. 1D).
Laboratory methods                                              It suggested that MDSCs may play important roles in
  Further details on Flow cytometry analysis, H&E, and          tumor recurrence after liver transplantation for HCC.
immunostaining, qRT-PCR, MDSC transwell assay, RT2
profiler PCR array can be found in the Supplementary             Higher CXCL10/TLR4 levels in small-for-size grafts were
methods section.                                                associated with tumor recurrence in HCC patients post-
                                                                liver transplantation
Statistical analysis                                              The intragraft levels of CXCL10 (p = 0.0191) and TLR4
  Categorical data were analyzed by Chi-square/Fisher’s         (p = 0.0091) were significantly higher in patients with small-
exact test. Comparison of continuous variables was per-         for-size grafts in contrast with large grafts at 2 h after
formed by Student’s t-test/Mann–Whitney U test where            reperfusion (Fig. 2A). Consistent with the results of MDSC
appropriate. Clinical survival analysis was implemented         analysis, the intragraft expressions of CXCL10 (p = 0.0068)
using Kaplan–Meier test. Data were presented as mean ±          and TLR4 (p = 0.0056) were also significantly up-regulated
SEM (standard error of the mean). Degrees of statistical        in the patients with HCC recurrence in contrast with those
significance was indicated using standardized asterisk           without HCC recurrence (Fig. 2B). Importantly, the intra-
nomenclature (*p < 0.05, **p < 0.01, ***p < 0.001). All         graft expressions of CXCL10 and TLR4 were highly corre-
analyses were performed with SPSS18.0 (SPSS, IL, USA)           lated (p = 0.0019, Fig. 2C). The co-localization of TLR4 and

Official journal of the Cell Death Differentiation Association
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489                                                                                    Page 4 of 14

Table 1 Comparison of baseline characteristics between                             circulatory MDSCs were significantly increased in the
non-recurrent and recurrent HCC patients after liver                               recipient rats with small-for-size graft compared to whole
transplantation.                                                                   graft post-transplantation (Day 1: p = 0.015, Day 3: p =
                              Non-recurrence         Recurrence          p value   0.038, Day 5: p = 0.043, Fig. 3A). Consistently, the intra-
                              (N = 281)              (N = 50)                      graft MDSCs (CD11b/c positive) were significantly
Gender (n)                                                               0.095     increased in the rats implanted with small-for-size liver
 Male                         225                    45                            graft (p = 0.0002, Fig. 3B). Furthermore, the intragraft
 Female                       56                     5                             levels of CXCL10 (Day 3: p = 0.0315), TLR4 (Day 3: p =
Age (years)                   55.38 ± 0.475          52.14 ± 1.146       0.006*    0.022, Day 5: p = 0.048), and CXCR3 (Day 5: p = 0.0043)
HBsAg (n)                                                                0.01*     were significantly higher in the small-for-size liver graft
 Negative                     61                     3                             (Fig. 3C).
 Positive                     220                    47                              Higher intratumoral microvessel density (MVD) (CD31
AFP (ng/ml)                   477.88 ± 164.67        4882.72 ± 2509.78   0.000*
                                                                                   positive) was found in small-for-size graft with tumor
No. of tumors (n)a            1.94 ± 0.10            3.34 ± 0.43         0.002*
                                                                                   recurrence model (Fig. 3D). This finding was consistent
The largest size of tumor     3.02 ± 0.08            4.15 ± 0.43         0.003*
(cm)a                                                                              with our previous studies30,31. The levels of CXCL10 in
Macrovascular invasion                                                   0.000*    tumor/non-tumor tissue (p = 0.0148, p = 0.0111) and
(n)a                                                                               TLR4 (p = 0.0343) in non-tumor tissue, but not CXCR3,
 No                           264                    39
                                                                                   were significantly higher in small-for-size graft than whole
 Yes                          7                      9
                                                                                   graft post-transplantation with tumor recurrence (Fig. 3E).
TNM staging (n)a                                                         0.000*
                                                                                   Furthermore, the intragraft MDSCs (CD11b/c positive)
 I                            109                    9
 II                           144                    25
                                                                                   were significantly increased in small-for-size grafts with
 III                          14                     13                            tumor recurrence (p = 0.0262, Fig. 3F).
Milan criteria (n)a                                                      0.000*
 Within criteria              192                    20                            CXCL10/TLR4 deficiency decreased mobilization and
 Beyond criteria              81                     29                            recruitment of monocytic MDSCs in the mouse IRH model
UCSF criteria (n)a                                                       0.000*      The direct role of CXCL10 signaling on MDSC mobili-
 Within criteria              222                    21                            zation was explored in the mouse IRH model using
 Beyond criteria              51                     28                            CXCL10−/−, TLR4−/− and CXCR3−/− mice, respectively.
Vascular permeation (n)a                                                 0.000*
                                                                                   The circulatory monocytic MDSCs were significantly
 No                           190                    18
                                                                                   reduced in CXCL10−/− (Day 1: p = 0.0108, Day 3: p =
 Yes                          71                     29
                                                                                   0.0351) and TLR4−/− (Day 1: p = 0.0106) mice. Further-
Tumor differentiation (n)a                                               0.025*
 Well-differentiated          78                     5
                                                                                   more, the hepatic monocytic MDSCs were obviously
 Moderately                   156                    33                            decreased in CXCL10−/− (Day 1: p = 0.0015) and TLR4−/−
 differentiated                                                                    (Day 1: p = 0.0021, Day 3: p = 0.0122) mice in contrast
 Poorly differentiated        12                     5                             with wild type ones after IRH (Fig. 4A, B). Interestingly,
 Undifferentiated             2                      0                             knockout of CXCL10 and TLR4 had no effects on the
Graft weight to recipient                                                0.018*
ESLV (%) (n)a
                                                                                   mobilization and recruitment of granulocytic MDSCs
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489                                                                                               Page 5 of 14

  Fig. 1 Increased MDSCs in HCC patients with small-for-size graft or tumor recurrence accompanied by poor disease-free survival post-
  transplantation. The circulatory and intragraft MDSCs were detected on Day 7 and hour 2 after liver transplantation, respectively. A The patients who
  received small-for-size liver grafts (GWR < 60%) had poor disease-free survival than patients with large liver grafts (GWR ≥ 60%) (n = 331). B Increased
  circulatory and intragraft MDSCs in patients with small-for-size grafts compared to those with large liver grafts (n = 61). C More circulatory and
  intragraft MDSCs in patients experienced HCC recurrence in contrast with non-recurrent patients (n = 61). D Poor disease-free survival in HCC patients
  with high MDSCs post-transplantation (n = 61). Scale bars: 100 μm. Error bars indicate standard error of mean; *p < 0.05, **p < 0.01. MDSC myeloid-
  derived suppressor cell, HCC hepatocellular carcinoma, GWR graft weight ratio.

Official journal of the Cell Death Differentiation Association
Monocytic MDSC mobilization promotes tumor recurrence after liver transplantation via CXCL10/ TLR4/MMP14 signaling
Liu et al. Cell Death and Disease (2021)12:489                                                                                          Page 6 of 14

  Fig. 2 Association of higher CXCL10/TLR4 levels in small-for-size graft with tumor recurrence in HCC patients after transplantation.
  Intragraft CXCL10 and TLR4 expressions were examined at hour 2 after liver transplantation. A Enhanced expression of CXCL10 and TLR4 in patients
  who received small-for-size liver grafts (GWR < 60%) compared to large liver grafts (GWR ≥ 60%) (n = 85). B Higher levels of CXCL10 and TLR4 in the
  recipients with HCC recurrence than non-recurrent patients (n = 85). C The significant positive correlation between intragraft CXCL10 and TLR4
  expression (n = 85). D Co-localization of TLR4 and MDSC (CD33 positive) in liver graft post-transplantation (a representative from three times
  independent experiments). Scale bars: 5 μm. Error bars indicate standard error of mean; *p < 0.05, **p < 0.01. HCC hepatocellular carcinoma, MDSC
  myeloid-derived suppressor cell, GWR graft weight ratio.

wild-type and TLR4−/− mice. The purity of isolated cells                     CXCL10 (300 ng/ml) treatment. The alteration of motility
was verified by flow cytometry (Supplementary Fig. 2A, B).                     genes and change folds were shown in Supplementary
RT2 profiler PCR array of mouse motility genes was per-                       Fig. 3A, B. Among the up-regulated genes in MDSCs with
formed from wild type and TLR4−/− MDSCs with/without                         CXCL10 administration (wild type + CXCL10 vs. wild type

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                                   Page 7 of 14

  Fig. 3 Association among MDSC mobilization, CXCL10/TLR4 expressions, and liver tumor recurrence in rat orthotopic liver transplantation
  models. A–C Rat orthotopic liver transplantation model. A More circulatory MDSCs in rats at Day 1, 3, 5 after small-for-size graft transplantation in
  contrast with whole graft. B Increased intragraft MDSCs (CD11b/c positive) in rats with small-for-size grafts at Day 5. C Increased intragraft mRNA
  expressions of CXCL10 (Day 3), TLR4 (Day 3, 5), CXCR3 (Day 5) in small-for-size liver graft. D–F Rat orthotopic liver transplantation with tumor
  recurrence model. D Higher MVD in the tumors from a small-for-size liver graft on Day 14 after transplantation. E Higher expressions of CXCL10 (both
  in tumor and non-tumor tissue) and TLR4 (in non-tumor tissue) in small-for-size liver graft with tumor recurrence. F More infiltrated MDSCs in small-
  for-size liver graft with tumor recurrence. N = 5/group; Scale bars: 50 μm; error bars indicate standard error of mean; *p < 0.05, **p < 0.01, ***p < 0.001.
  MDSC myeloid-derived suppressor cell, MVD microvessel density.

MDSCs), MMP14 was increased most with 4.53 folds. It                              also decreased most with −16.62 folds compared to wild-
indicated that CXCL10 up-regulated MMP14 to increase                              type ones with CXCL10 treatment (Fig. 5A, Supplementary
the mobilization of MDSCs. Interestingly, among the                               Fig. 3B). It demonstrated that MMP14 promoting MDSC
down-regulated genes in TLR4−/− MDSCs, MMP14 was                                  mobilization was TLR4 dependent.

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                          Page 8 of 14

  Fig. 4 Reduced mobilization and recruitment of monocytic MDSCs by knockout of CXCL10/TLR4 in mouse IRH model. A Less circulatory and
  hepatic monocytic MDSCs in CXCL10−/− mice compared to wild-type ones. B Significantly decreased mobilization and recruitment of monocytic
  MDSCs in TLR4−/− mice. C No significant change of circulatory and hepatic monocytic MDSCs by knockout of CXCR3. D Reduced TLR4+ monocytic
  MDSCs in the circulation and liver of CXCL10−/− mice compared to wild-type ones after IRH. N = 5/group; Error bars indicate Standard Error of Mean;
  *p < 0.05, **p < 0.01. MDSC myeloid-derived suppressor cell, IRH ischemia/reperfusion plus major hepatectomy, WT wild type.

  In the transwell assay, recombinant CXCL10 obviously                       from the wild type MDSCs, there was no significant
recruited more wild type MDSCs to the bottom wells in a                      change of migration for TLR4−/− MDSCs and MMP14
dosage-dependent manner (100 ng/ml vs. no CXCL10:                            antibody blocking wild type MDSCs (30 μg/ml) with or
p = 0.0075, 300 ng/ml vs. no CXCL10: p = 0.0002,                             without recombinant CXCL10 addition (Fig. 5B). The
300 ng/ml vs. 100 ng/ml CXCL10: p = 0.0005). Different                       result of CXCR3−/− MDSCs was similar to the wild-type

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                          Page 9 of 14

  Fig. 5 MDSC mobilization regulated by CXCL10/TLR4 through MMP14 and higher levels of TLR4 on monocytic MDSCs. A MMP14, as the
  most up-regulated gene, in wild type MDSCs with CXCL10 addition compared to those without treatment; MMP14, as the most down-regulated
  gene, in TLR4−/− MDSCs compared to wild type ones with the same CXCL10 treatment by the screen of mouse motility genes. B More wild type
  MDSCs transferred to the bottom well in a CXCL10 dosage-dependent manner, while no significantly changed numbers of TLR4−/− MDSCs and
  MMP14 Ab-blocking wild type MDSCs (30 μg/ml) in the bottom well with/without CXCL10 addition. C Augmentation of TLR4 on monocytic MDSCs
  in contrast with granulocytic MDSCs in mouse IRH model (N = 6/group). Scale bars: 100 μm. Error bars indicate standard error of mean; *p < 0.05,
  **p < 0.01, ***p < 0.001. MDSC myeloid-derived suppressor cell, MMP14 matrix metallopeptidase 14, Ab antibody, IRH ischemia/reperfusion plus major
  hepatectomy, WT wild type.

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                Page 10 of 14

ones. It suggested that CXCR3 deficiency or not did not          indicated that the patients implanted with small-for-size
influence the MDSC migration induced by CXCL10                   graft had higher HCC recurrence and poor disease-free
(Supplementary Fig. 4A). Taken together, CXCL10/TLR4            survival. These results were consistent with the previous
mobilized MDSCs through MMP14.                                  clinical experience from both Western and Eastern
  The reason why more monocytic MDSCs, but not gran-            cohorts4,5,32. Our previous animal studies also demon-
ulocytic MDSCs, recruited to the liver after reperfusion was    strated the significance of acute-phase small-for-size graft
further explored. The TLR4 expression on monocytic              injury on tumor growth and invasiveness after liver
MDSCs was found significantly higher compared to granu-          transplantation30. However, the mechanism of small-for-
locytic MDSCs (p < 0.0001, Fig. 5C, Supplementary Fig. 4B)      size liver graft injury leading to HCC recurrence is still
in mice. These findings indicated that more monocytic            largely unknown.
MDSCs were mobilized to the liver because of higher TLR4           We first identified that MDSCs recruited by CXCL10/
expression, which was the receptor of CXCL10.                   TLR4 during acute phase inflammation played a critical
                                                                role in tumor recurrence after liver transplantation
Cancer recurrence was inhibited with reduced monocytic          through a series of clinical analyses, animal models, and
MDSCs by knockout or inhibition of CXCL10/TLR4 in the           in vitro functional experiments. In addition to the
mouse hepatic IRH with tumor recurrence model                   immunosuppressive properties, MDSCs also possess
  (1) In CXCL10−/− and TLR4−/− mice                             protumorigenic functions such as angiogenesis, che-
  To explore the direct role of CXCL10/TLR4 signaling on        moresistance, and metastasis33–35. According to our
MDSC mobilization and late tumor recurrence, mouse              results, more MDSCs were accumulated in the recipients
hepatic IRH with tumor recurrence model was applied.            with small-for-size grafts during acute injury. The early
Compared to wild-type mice, knockout of CXCL10 or               recruitment of MDSCs into the liver graft may provide a
TLR4 significantly inhibited tumor development. However,         favorable environment for tumor recurrence. In the late
CXCR3 deficiency could not obviously reduce the tumor            phase, the recipients with small-for-size graft had more
burden with no significant change of CD31 positive cells in      tumor recurrence with increased MDSCs compared to the
tumor tissue (Fig. 6A, B). Simultaneously, the circulatory      whole graft. These findings indicated that MDSCs traf-
and hepatic monocytic MDSCs were significantly                   ficking to the liver graft might be detained and pro-
decreased in CXCL10−/− (circulatory: p = 0.0098; hepatic:       liferated, and promote the tumor recurrence. Recently,
p = 0.0213) and TLR4−/− (circulatory: p = 0.0134; hepatic:      MDSCs were proposed to be applied for attenuation of
p = 0.0247) mice of IRH with tumor recurrence model             allograft rejection based on their immunosuppressive
(Fig. 6C). Consistent with our in vitro studies, MMP14          function36. Our studies may raise the caution for using
expression was also decreased in TLR4−/− mice (p <              MDSCs in transplantation for cancer patients due to their
0.0001, Fig. 6D). Although the circulatory granulocytic         role in tumor recurrence. The balance between decreasing
MDSCs were decreased by knockout of CXCL10 or TLR4,             allograft rejection and avoiding cancer recurrence for
the population of hepatic granulocytic MDSCs was not            HCC recipients should be considered.
changed (Supplementary Fig. 5A).                                   The recruitment of MDSCs to liver graft by CXCL10 is
  (2) TLR4 inhibition                                           through its receptor. Intriguingly, only knockout of TLR4,
  TLR4 inhibitor (CLI095) was used to explore its ther-         instead of CXCR3, could obviously decrease the MDSC
apeutic effects in tumor recurrence post IRH. Wild type         mobilization at the early stage and reduce tumor pro-
mice with CLI095 treatment had a smaller tumor burden           gression at the late phase. In addition, the role of CXCL10
with less CD31 expression (MVD) in contrast with non-           mobilizing MDSCs via TLR4 was further validated by the
treated mice (Fig. 7A, B). Consistently, hepatic monocytic      high correlation of intragraft CXCL10 and TLR4, co-
MDSCs were significantly reduced by CLI095 treatment             localization of TLR4 and MDSC marker in patients, the
compared to the non-treatment group (p = 0.0149) while          decrease of TLR4+ MDSCs in CXCL10−/− mice and our
no significant change for the circulatory monocytic              in vitro studies. TLR4, as the pathogen recognition
MDSCs (Fig. 7C). Hepatic MMP14 level was also reduced           receptor located on the cell membrane, is responsible for
in the CLI095 treatment group (p = 0.0001, Fig. 7D). The        the liver ischemia/reperfusion injury after transplanta-
circulatory and hepatic granulocytic MDSCs were not             tion10. In our mouse tumor recurrent model, TLR4
obviously altered between the two groups (Supplementary         inhibitors could effectively reduce hepatic MDSCs and
Fig. 5B).                                                       inhibit tumor growth. Therefore, targeting TLR4, could
                                                                not only protect the liver graft from ischemia/reperfusion
Discussion                                                      injury but also reduce MDSC mobilization to decrease
  Tumor recurrence is a critical issue that affects the         tumor recurrence after transplantation.
outcomes of liver transplantation for HCC patients. In             MMP14, a component of MMP family37, was identified
this study, our clinical analysis of 331 HCC recipients         as the most up-regulated motility gene by CXCL10

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                            Page 11 of 14

  Fig. 6 Inhibited cancer recurrence with reduced monocytic MDSCs and MMP14 in CXCL10−/− or TLR4−/− mice of IRH with tumor
  recurrence model. A Inhibited tumor development by knockout of CXCL10 or TLR4; scale bars: 500 μm. B No significant change of infiltrated CD31
  positive cells in tumor tissue between wild type and CXCR3−/− mice; Scale bars: 50 μm. C Less circulatory and hepatic monocytic MDSCs in CXCL10−/−
  or TLR4−/− mice. D Significant decrease of infiltrated MMP14 in TLR4−/− mice in contrast with wild-type ones; scale bars: 100 μm. N = 5/group; error
  bars indicate standard error of mean; *p < 0.05, **p < 0.01, ***p < 0.001. MDSC myeloid-derived suppressor cell, MMP14 matrix metallopeptidase 14, IRH
  ischemia/reperfusion plus major hepatectomy, WT wild type.

addition, while the most down-regulated one in TLR4−/−                         Nevertheless, the regulatory role of MMP14 on MDSCs’
MDSCs. These findings indicated that MMP14 was the                              motility was not clear. Our study demonstrated that the
most critical molecule in MDSC migration by CXCL10/                            mobilization of MDSCs was mediated by MMP14, which
TLR4 signaling. Overexpressed MMP14 has been repor-                            was regulated by CXCL10/TLR4. It provided a new
ted in tumor-residing MDSCs, which facilitate tumor cell                       mechanism of MDSC motility. Thus, the overexpression
invasion and metastasis of mammary carcinoma35.                                of MMP14 not only facilitated tumor cell invasion but

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                          Page 12 of 14

  Fig. 7 Decrease of cancer recurrence, monocytic MDSCs, and MMP14 by TLR4 inhibition in mouse liver IRH with tumor recurrence model.
  A Reduced tumor progression by the treatment of TLR4 inhibitors (CLI095) compared to no treatment wild type mice; scale bars: 500 μm. B The
  obvious decline of infiltrated CD31 positive cells in tumor tissue by TLR4 inhibition; scale bars: 50 μm. C Decreased liver monocytic MDSCs by TLR4
  inhibition. D Less infiltrated MMP14 in the mice with the treatment of TLR4 inhibitors; scale bars: 100 μm. E Research summary: monocytic MDSCs
  were mobilized and recruited to liver graft through CXCL10/TLR4/MMP14 signaling during acute phase injury, and to promote HCC recurrence after
  transplantation. N = 5/group; error bars indicate standard error of mean; *p < 0.05, **p < 0.01, ***p < 0.001. MDSC myeloid-derived suppressor cell,
  MMP14 matrix metallopeptidase 14, IRH ischemia/reperfusion plus major hepatectomy, WT wild type.

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                                                    Page 13 of 14

also mobilized MDSCs to the liver graft promoting HCC                                        Ethics statement
recurrence after transplantation. Our studies further                                        All clinical record access and tissue sample collection were approved with
                                                                                             signed consent forms by each donor and recipient. This study was approved
certified that deficiency of CXCL10/TLR4 or TLR4 inhi-                                         by the Committee on the Use of Live Animals in Teaching and Research
bitors could effectively reduce the monocytic MDSC                                           (CULATR), The University of Hong Kong.
recruitment with decreased MMP14, leading to the
                                                                                             Conflict of interest
smaller tumor occupation.                                                                    The authors declare no competing interests.
  Monocytic MDSCs, rather than granulocytic MDSCs,
were found to be mobilized by CXCL10/TLR4 after
hepatic ischemia-reperfusion. This might be explained by                                     Publisher’s note
                                                                                             Springer Nature remains neutral with regard to jurisdictional claims in
the higher expression of TLR4 on monocytic MDSCs in                                          published maps and institutional affiliations.
contrast to granulocytic ones. Consistently, there were
more monocytic MDSCs accumulated in the mice with                                            Supplementary information The online version contains supplementary
                                                                                             material available at https://doi.org/10.1038/s41419-021-03788-4.
larger tumor volume. Monocytic and granulocytic
MDSCs are the main subsets of MDSCs. Unlike termin-
                                                                                             Received: 16 January 2021 Revised: 3 May 2021 Accepted: 3 May 2021
ally differentiated granulocytic MDSCs, monocytic
MDSCs can differentiate into dendritic cells and macro-
phages18. The monocytic MDSCs could be one of the
major sources for tumor-associated macrophages (M2                                           References
macrophages) with immunosuppressive properties. M2                                            1. Villanueva, A., Hernandez-Gea, V. & Llovet, J. M. Medical therapies for hepa-
                                                                                                  tocellular carcinoma: a critical view of the evidence. Nat. Rev. Gastroenterol.
macrophages have been reported to promote HCC inva-                                               Hepatol. 10, 34 (2013).
siveness and recurrence in our previous study38. In                                           2. Bray, F. et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence
addition, the monocytic MDSCs increase NO through                                                 and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 68,
                                                                                                  394–424 (2018).
STAT1/inducible nitric oxide synthase signaling. NO can                                       3. Man, K. Recurrent malignancy:are we pushing the envelope?. Liver Transpl 23,
suppress T-cell functions through various mechanisms to                                           S81–S84 (2017).
facilitate tumor progression18. Although the direct role of                                   4. Fisher, R. et al. Hepatocellular carcinoma recurrence and death following
                                                                                                  living and deceased donor liver transplantation. Am. J. Transpl. 7, 1601–1608
MDSCs on HCC invasiveness was not focused on in the                                               (2007).
current study, we did provide evidence that the acute                                         5. Kulik, L. et al. Outcomes of living and deceased donor liver transplant reci-
phase inflammation of graft injury might initiate mono-                                            pients with hepatocellular carcinoma: results of the A2ALL cohort. Am. J.
                                                                                                  Transpl. 12, 2997–3007 (2012).
cytic MDSC accumulation, which altered the liver                                              6. Van Der Bilt, J. D. et al. Ischemia/reperfusion accelerates the outgrowth of
microenvironment favoring tumor recurrence in the                                                 hepatic micrometastases in a highly standardized murine model. Hepatology
scenario of liver transplantation.                                                                42, 165–175 (2005).
                                                                                              7. Oldani, G. et al. Pre-retrieval reperfusion decreases cancer recurrence after rat
  Taken together, the monocytic MDSCs were mobilized                                              ischemic liver graft transplantation. J. Hepatol. 61, 278–285 (2014).
and recruited to the liver graft through CXCL10/TLR4/                                         8. Brownell, J. & Polyak, S. J. Molecular pathways: hepatitis C virus, CXCL10, and
MMP14 signaling during acute phase injury, leading to                                             the inflammatory road to liver cancer. Clin. Cancer Res. 19, 1347–1352 (2013).
                                                                                              9. Hintermann, E., Bayer, M., Pfeilschifter, J. M., Luster, A. D. & Christen, U. CXCL10
HCC recurrence post-transplantation through our inte-                                             promotes liver fibrosis by prevention of NK cell mediated hepatic stellate cell
grated study with clinical analyses, animal models, and                                           inactivation. J. Autoimmun. 35, 424–435 (2010).
in vitro experiments. Targeting CXCL10/TLR4/MMP14                                            10. Zhai, Y., Petrowsky, H., Hong, J. C., Busuttil, R. W. & Kupiec-Weglinski, J. W.
                                                                                                 Ischaemia–reperfusion injury in liver transplantation—from bench to bedside.
to inhibit MDSC mobilization will be critical in the                                             Nat. Rev. Gastroenterol. Hepatol. 10, 79 (2013).
devising of novel therapeutic strategies against HCC                                         11. Datta, D. et al. Ras-induced modulation of CXCL10 and its receptor splice
recurrence after transplantation.                                                                variant CXCR3-B in MDA-MB-435 and MCF-7 cells: relevance for the devel-
                                                                                                 opment of human breast cancer. Cancer Res. 66, 9509–9518 (2006).
                                                                                             12. Zipin-Roitman, A. et al. CXCL10 promotes invasion-related properties in
Acknowledgements                                                                                 human colorectal carcinoma cells. Cancer Res. 67, 3396–3405 (2007).
We would like to thank Dr. Alice Yuen Yuen Ma for her contribution to the flow                13. Wightman, S. et al. Oncogenic CXCL10 signalling drives metastasis develop-
cytometry experiments.                                                                           ment and poor clinical outcome. Br. J. Cancer 113, 327 (2015).
                                                                                             14. Schulthess, F. T. et al. CXCL10 impairs β cell function and viability in diabetes
                                                                                                 through TLR4 signaling. Cell Metab. 9, 125–139 (2009).
Author contributions
                                                                                             15. Sahin, H. et al. Proapoptotic effects of the chemokine, CXCL 10 are mediated
H.L. and W.Y. designed the study; H.L., C.L., W.Y., L.P., J.L., J.Z., W.Z., X.L., T.N. and
                                                                                                 by the noncognate receptor TLR4 in hepatocytes. Hepatology 57, 797–805
X.Y. performed the experiments and interpretation of data; K.M., C.L. obtained
                                                                                                 (2013).
fundings and supervised the study; All authors were involved in technical
                                                                                             16. Orci, L. A. et al. Effects of the gut–liver axis on ischaemia-mediated hepato-
support and writing/revising the manuscript. All authors had final approval of
                                                                                                 cellular carcinoma recurrence in the mouse liver. J. Hepatol. 68, 978–985 (2018).
the submitted version.
                                                                                             17. Ilkovitch, D. & Lopez, D. M. The liver is a site for tumor-induced myeloid-
                                                                                                 derived suppressor cell accumulation and immunosuppression. Cancer Res.
Funding                                                                                          69, 5514–5521 (2009).
This work was supported by the General Research Funding (GRF:                                18. Gabrilovich, D. I. & Nagaraj, S. Myeloid-derived suppressor cells as regulators of
17124219;17122517;17159616; 17115515), Theme-based Research Scheme                               the immune system. Nat. Rev. Immunol. 9, 162 (2009).
(TRS: T12-703/19R), and Collaborative Research Fund (C7026-18GF) from the                    19. Liu, C. et al. Expansion of spleen myeloid suppressor cells represses NK cell
Research Grant Council, Hong Kong.                                                               cytotoxicity in tumor-bearing host. Blood 109, 4336–4342 (2007).

Official journal of the Cell Death Differentiation Association
Liu et al. Cell Death and Disease (2021)12:489                                                                                                               Page 14 of 14

20. Huang, B. et al. Gr-1+ CD115+ immature myeloid suppressor cells mediate                29. Cheng, Q. et al. Distinct mechanism of small-for-size fatty liver graft injury-
    the development of tumor-induced T regulatory cells and T-cell anergy in                   wnt4 signaling activates hepatic stellate cells. Am. J. Transpl. 10, 1178–1188
    tumor-bearing host. Cancer Res. 66, 1123–1131 (2006).                                      (2010).
21. Sinha, P., Clements, V. K., Bunt, S. K., Albelda, S. M. & Ostrand-Rosenberg, S.        30. Man, K. et al. The significance of acute phase small-for-size graft injury on
    Cross-talk between myeloid-derived suppressor cells and macrophages sub-                   tumor growth and invasiveness after liver transplantation. Ann. Surg. 247,
    verts tumor immunity toward a type 2 response. J. Immunol. 179, 977–983                    1049–1057 (2008).
    (2007).                                                                                31. Man, K. et al. Molecular signature linked to acute phase injury and tumor
22. Chiu, D. K.-C. et al. Hypoxia inducible factor HIF-1 promotes myeloid-derived              invasiveness in small-for-size liver grafts. Ann. Surg. 251, 1154–1161 (2010).
    suppressor cells accumulation through ENTPD2/CD39L1 in hepatocellular                  32. Lo, C. et al. Living donor versus deceased donor liver transplantation for early
    carcinoma. Nat. Commun. 8, 517 (2017).                                                     irresectable hepatocellular carcinoma. Br. J. Surg. 94, 78–86 (2007).
23. Chiu, D. K. C. et al. Hypoxia induces myeloid-derived suppressor cell recruit-         33. Yang, L. et al. Expansion of myeloid immune suppressor Gr+ CD11b+ cells in
    ment to hepatocellular carcinoma through chemokine (C-C motif) ligand 26.                  tumor-bearing host directly promotes tumor angiogenesis. Cancer Cell 6,
    Hepatology 64, 797–813 (2016).                                                             409–421 (2004).
24. Ehling, J. & Tacke, F. Role of chemokine pathways in hepatobiliary cancer.             34. Bruchard, M. et al. Chemotherapy-triggered cathepsin B release in myeloid-
    Cancer Lett. 379, 173–183 (2016).                                                          derived suppressor cells activates the Nlrp3 inflammasome and promotes
25. Man, K. et al. Graft injury in relation to graft size in right lobe live donor             tumor growth. Nat. Med. 19, 57 (2013).
    liver transplantation: a study of hepatic sinusoidal injury in correlation with        35. Yang, L. et al. Abrogation of TGFβ signaling in mammary carcinomas recruits
    portal hemodynamics and intragraft gene expression. Ann. Surg. 237, 256                    Gr-1+ CD11b+ myeloid cells that promote metastasis. Cancer Cell 13, 23–35
    (2003).                                                                                    (2008).
26. Ling, C.-C. et al. Post-transplant endothelial progenitor cell mobilization via        36. Nakamura, T. & Ushigome, H. Myeloid-derived suppressor cells as a regulator
    CXCL10/CXCR3 signaling promotes liver tumor growth. J. Hepatol. 60,                        of immunity in organ transplantation. Int J. Mol. Sci. 19, 2357 (2018).
    103–109 (2014).                                                                        37. Lu, H. et al. KLF8 and FAK cooperatively enrich the active MMP14 on the cell
27. Ye, D. et al. Toll-like receptor-4 mediates obesity-induced non-alcoholic stea-            surface required for the metastatic progression of breast cancer. Oncogene 33,
    tohepatitis through activation of X-box binding protein-1 in mice. Gut 61,                 2909–2917 (2014).
    1058–1067 (2012).                                                                      38. Yeung, O. W. et al. Alternatively activated (M2) macrophages promote tumour
28. Man, K. et al. Liver transplantation in rats using small-for-size grafts: a study of       growth and invasiveness in hepatocellular carcinoma. J. Hepatol. 62, 607–616
    hemodynamic and morphological changes. Arch. Surg. 136, 280–285 (2001).                    (2015).

Official journal of the Cell Death Differentiation Association
You can also read